Overview
server_permanent_offline_time specifies the duration after which a node heartbeat is considered permanently lost. When a node's heartbeat is lost for this period, the Root Service processes the data replicas on that OBServer node, removes it from the Paxos member group, and replenishes the data on other available OBServer nodes to maintain the integrity of the Paxos member group.
Privilege requirements
Query parameters
Both the
systenant and all user tenants can query this parameter using theSHOW PARAMETERSstatement or theGV$OB_PARAMETERSview.Modify parameters
Only the
systenant can modify this parameter. User tenants cannot modify it.
Attributes
Attribute |
Description |
|---|---|
| Parameter type | Time |
| Default Value | 3600s |
| Value range | 20s, +∞) |
| Modifiable | Yes, you can use theALTER SYSTEM SETStatement modification. |
| Whether to restart the OBServer node for the change to take effect | No, the setting takes effect immediately. |
Usage instructions
By setting this parameter, unnecessary replica replication can be avoided. Data replica replication on the permanently offline node is initiated only after the replica loss duration exceeds the specified value.
The applicable scenarios and recommended values for this parameter are as follows:
OceanBase AI Database version upgrade scenarios: It is recommended to set the value of this parameter to
72h.Note
When upgrading OceanBase AI Database to OCP V2.3.5 or later, the system automatically sets the value of this parameter to
72h.Hardware replacement scenario for OBServer nodes: It is recommended to set the value of this parameter to
4h.OBServer node clearance and online scenario: It is recommended to set the value of this parameter to
10mto enable the cluster to go online quickly.
Configuration example
Set the node heartbeat interruption time threshold to 3600s.
obclient> ALTER SYSTEM SET server_permanent_offline_time = '3600s';
References
For more information about Root Service, see [Database layer high availability].